Note: Descriptions are shown in the official language in which they were submitted.
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
METHOD AND APPARATUS FOR DETERMINING
TARGET DEPTH. BRIGHTNESS AND SIZE WITHIN A BODY REGION
Field of the invention
This invention relates to an apparatus and method for fields of non-
invasive medical imaging, medical research, pathology, and drug discovery and
development.
Background of the invention
In a variety of medical diagnostic and therapeutic settings, as well as in
biomedical research, it is desirable to image a subsurface target or area
within a
body region of a subject. For example, non-invasive locating and imaging of
part
or all of a solid tumor, an area of myocardial ischemia, the distribution of a
therapeutic compound administered to the subject, or the progression of a
disease may provide useful research or diagnostic information. Ideally, an
imaging method is able to locate within a body region a target of interest,
and
provide information about the target's shape, size, number of cells, and depth
below the surface of the body region. However, until now, methods that have
been used and/or proposed for subsurface body target imaging have generally
been limited to those using ionizing radiation such as X-rays, expensive and
bulky equipment such as Magnetic Resonance imaging (MRI), or ultra-sound.
X-rays have excellent tissue penetration, and when used in conjunction
with computed tomography (CT) or computed axial tomography (CAT), can
produce superior image quality. However, X-rays have limited use in monitoring
disease progress because exposure to X-rays is potentially harmful if such
_ exposure is prolonged. X-rays can.be used to locate and image compositions
that have localized at a target within a body region, but always with exposure
to
the potential harm associated with X-ray radiation. X-rays, however, cannot
readily be used to image the expression of gene products in vivo and determine
the depth and/or shape of a target expressing such gene products.
MRI is also an excellent method for imaging targets, areas, and structures
in a body region of a subject. Although MRI is not thought to possess harmful
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
attributes like those associated with ionizing radiation, the expense and
bulky
equipment size needed to use MRI make it impractical for many applications or
situations. MRI can provide two and three-dimensional information about
targets
within a body region of a subject, but is less effective at imaging
physiological
activity associated with a target.
Ultrasound or ultrasonography is the use of high-frequency sound
(ultrasonic) waves to produce images of structures within the human body.
Ultrasonic waves are sound waves that are above the range of sound audible to
humans. Ultrasonic waves are produced by the electrical stimulation of a
piezoelectric crystal and such waves can be aimed at specific body regions. As
the waves travel through body tissues within a body region, they are reflected
back at any point where there is a change in tissue density, as, for instance,
in
the border between two different organs of the body. Ultrasound offers the
advantages of not using radiation or radioactive material, and employs lesser
expensive and less bulky equipment than MRI, but is limited to only discerning
differences in density of underlying tissue and structures. Accordingly,
ultrasound cannot effectively track and monitor the progress of an infection
unless such infection results in a discernable shift in density of the target
tissue.
Ultrasound cannot image or detect the physiological functions of tissues or
organs.
Until now, Positron Emission Tomography or P.E.T. was unique among
imaging techniques because it produces an image of organ or tissue function.
Other imaging techniques such as X-ray, CT, MRI, and sonography depict organ
or tissue anatomy but cannot discern physiological activity within them. To
image a specific biochemical activity of an organ, a radioactive substance,
called
a radiotracer or radiopharmaceutical, is injected into the body or inhaled.
The
. tracer is usually a radioactive equivalent of a substance that occurs
naturally
within the body such as water or sugar. The radioactive isotope is identical
to the
body's own nonradioactive isotope except that its atoms have a different
number
of neutrons. Consequently, a subject's body is burdened with radioactive
material, and the potential harm associated with such material. P.E.T cannot
detect non-isotopic expression products from transgenic tissues, organs, or
2
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
transgenic organisms. Scintigraphy, a diagnostic technique in which a two-
dimensional picture of a bodily radiation source is obtained by the use of
radioisotopes, may also be used for imaging structures and their functions.
Scintigraphy, however, is not suitable for determining the depth of a target
in a
body region of a subject.
In view of the above-mentioned technologies for locating and imaging a
target in a body region of a subject, there is a need for methods and devices
to
determine the depth and/or the shape and/or number of cells of such target
without having to use radioactivity, radiation, or expensive and bulky
equipment.
The invention disclosed herein meets these needs.
Summary of the Invention
In one aspect, the invention includes a method of investigating the
location, size and number of cells, of a light-emitting source in a subject.
In
practicing the method, one initially obtains a first measured light intensity
profile
constructed by measuring, from a first perspective with a photodetector
device,
photons which (i) originate from the light-emitting source, (ii) travel
through turbidv
biological tissue of the subject, and (iii) are emitted from a first surface
region of .
interest of the subject. The light-intensity profile is matched against a
parameter-
based biophotonic function to estimate function parameters such as depth and
size. The parameters so determined are refined using data other than the first
measured light intensity profile, to obtain an approximate depth and size of
the
source in the subject. The additional data may be data measured from the
subject, data from modeling analysis, or data relating to the wavelength of
photons emitted from the surface of the subject. As examples:
The method typically includes generating a 2-D or 3-D visual
representation of the light-emitting source, using the approximate depth and
shape of the source in the subject, and superimposing the visual
representation
onto a 2-D or 3-D image of the subject.
The additional data may be obtained from computer simulation of the
diffusion of light from a light-emitting source in a turbid medium. One
preferred
simulation approach involves (i) generating a plurality of theoretical light-
intensity
3
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
profiles, based on a model of photon diffusion from a glowing source located
at
one of a plurality of depths, and having one of a plurality of sizes and
shapes,
through a turbid medium having absorption and scattering properties similar to
those of biological tissue, (ii) comparing the quality of fit between each of
the
plurality of theoretical light-intensity profiles and the first measured light
intensity
profile, (iii) selecting the theoretical light intensity profile which
provides to the
first measured light intensity profile, and (iv) obtaining an approximate
depth,
shape and brightness of the source in the subject using parameters from the
theoretical light intensity profile selected in (iii). The method may include
employing in a photon-scattering model, one or more predetermined tissue-
specific light-scattering coefficients corresponding to tissue through which
the
photons travel.
In another general embodiment, the additional data are obtained by
measuring light emission from the subject at two or more different wavelengths
between about 400 nm and about 1000 nm, determining the relative light
intensities measured at the different wavelengths, and comparing the
determined
relative light intensities with known relative signal intensities at the
different
wavelengths, as a function of tissue depth. Alternatively, the spectrum of
light
intensities is measured over a selected wavelength range between about 400-
1000 nm, and the measured spectrum is compared with a plurality of spectra
measured from light-emitting sources placed at various depths within tissue,
to
determine the depth of the light-emitting source from matching the measured
spectrum with the known spectra.
In various embodiments for a light-emission source, the source is a
luminescent moiety or a fluorescent moiety; the light-emitting source is
administered to the subject and binds to a selected target or is otherwise
localized
in the subject prior to the measuring; and the light-emitting source is a
light-
generating protein (e.g., a luciferase, a green fluorescent protein, etc.),
such as a
light-generating protein expressed by biological cells of the subject or
biological
cells administered to the subject.
In another aspect, the invention includes apparatus for use in
investigating the location and size of a light-emitting source in a subject.
The
4
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
apparatus includes a light-tight enclosure within which light-emission events
from
the subject can be detected and an optical system associated with the
enclosure
for use in generating a first light intensity profile constructed by
measuring, from
a first perspective within the enclosure, photons which (i) originate from the
light-
s emitting source, (ii) travel through turbid biological tissue of the
subject, and (iii)
are emitted from a first surface region of interest of the subject. A
computational
unit operatively connected to the photo-detector functions to (i) fit the
first
measured light intensity profile with a parameter-based biophotonic function;
and
(ii) refine the parameters of the biophotonic function, using data other than
the
first measured light intensity profile, to generate an approximate depth and
shape of the source in the subject.
The optical system preferably includes an intensified or cooled charge-
coupled-device (CCD), and a lens for focusing light onto the CCD. The optical
system may configured in such a way as to detect photons emitted from a
plurality of different selected surface regions of interest of the subject.
The
system may include one or more filters for transmitting photons within
different
selected wavelength ranges, e.g., above and below 600 nm.
The computational unit may include a data file of model biophotonic
functions representing light-emitting sources of various sizes at various
depths,
for curve matching with the first spatial profile.
Where the optical system includes filters for wavelength discrimination,
the computation unit is operable to carry out at least one of the parameter-
refinement operations:
(i) determining the relative light intensities measured at the different
wavelengths, and comparing the determined relative light intensities with
known
or calculated relative signal intensities at the different wavelengths, as a
function
of tissue depth; and
(ii) comparing the measured spectrum with a plurality of spectra
measured from light-emitting sources placed at various depths within tissue,
and
determining the depth of the light-emitting source from matching the measured
spectrum with the known spectra.
5
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
In another embodiment, the computational unit is operable to average the
intensity pattern image into a single intensity value, and further to
determine
source size by integrated light intensity values generated as a function of a
light-
emitting source of a particular size and shape.
The computational unit may have a database containing a plurality of
theoretical light-intensity profiles, based on a model of photon diffusion
from a
light-emitting source located at one of a plurality of depths, and having one
of a
plurality of sizes and shapes, through a turbid medium having absorption and
scattering properties similar to those of biological tissue. Here the
computational
unit is operable to (i) compare the quality of fit between each of the
plurality of
theoretical light-intensity profiles and first measured light intensity
profile, (ii)
select the theoretical light intensity profile which provides the best fit to
the first
measured light intensity profile, and (iii) obtain an approximate depth and
shape
of the source in the subject using parameters from the theoretical light
intensity
profile selected in (ii).
In addition, the computational unit is operable to generate a visual 2-, or
3-dimensional representation of the light-emitting source, using the
approximate
depth and shape of the source in the subject, and superimposing the visual
representation onto a 2- or 3-dimensional image of the subject.
In another aspect, the invention provides a method of determining the
depth of a light-emitting source in a subject. In practicing the method, the
light
emission intensity from the subject at two or more different wavelengths
between
about 400 and about 1000 nm is measured. The depth of the light-emitting
source is determined using the measured light intensities and information
related
to the optical properties of the subject's tissue, e.g., coefficients of
attenuation
and diffusion in the subject.
Information relating to optical properties, e.g., the coefficients of
attenuation and diffusion in the subject, may be obtained by direct
measurement
of the coefficients in material that is the same or similar to that of the
subject.
Alternatively, optical-properties information may be obtained indirectly, by
determining, at two or more different wavelengths, light intensities from a
light-
emitting source located at each of a plurality of depths in tissue or material
6
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
corresponding to that of the subject. The desired depth determination is then
carried out by matching the measured light intensities with the light
intensities
determined at each of the plurality of depths.
In the latter approach, the spectral profile of light intensities from the
light-
emitting source may be compared (matched) with each of a plurality of spectral
profiles of light intensities from a light-emitting source at each of a
plurality of
depths.
These and other objects and features of the invention will become more
fully apparent when the following detailed description of the invention is
read in
conjunction with the accompanying drawings.
Brief Description of the Drawings
Fig. 1 A is a cross-sectional view of an apparatus constructed in
accordance with the invention, for use in investigating the location and size
of a
light-emitting source in a subject;
Fig. 1 B illustrates features of the apparatus in Fig. 1 A in schematic form;
Fig. 2 is a flowchart of the general steps that can be carried out by the
apparatus, in practicing the method of the invention;
Figs. 3A is a surface light-intensity image from a light-emitting source in
an animal subject.
Fig. 3B is a light emission profile from the animal subject in Fig. 3A,
showing the curve fit between the measured emission profile (solid lines) and
a
light-emission profile calculated from a diffusion model;
Figs. 4A and 4B are plots of the emission spectra of bacterial luciferase in
vitro (solid line), and in vivo (dashed lines) at various depths in a mouse
thigh, as
indicated, where the plot in Fig. 4B has been scaled to show the peak
intensity in
vivo in the region between 600-700 nm;
Fig. 5 shows various locations in an animal at which transmission spectra
were obtained.
Figs. 6A and 6B show transmission spectra through a live experimental
animal at the various animal locations shown in Fig. 5;
7
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
Fig. 6C illustrates a double integrating sphere apparatus useful for
measuring optical properties of a sample, e.g., tissue;
Fig. 6D illustrates the experimental set up of a radial fiber probe, useful
for
measuring optical properties of tissue in vivo;
Figs. 7A and 7B are plots of the absorption coefficients;
Figs. 7C and 7D are plots of the isotropic or reduced scattering
coefficients of various tissues as a function of wavelength;
Fig. 8A illustrates components of the apparatus used for absolute
calibration determination;
Fig. 8B illustrates how absolute calibration allows determination of cell
count from an animal subject;
Figs. 9A and 9B illustrate how photons diffuse through tissue (9A) and
how the photons that diffuse to the surface of the tissue are captured in the
invention (9B);
Fig. 1 OA shows a Monte Carlo simulation following a photon through a
"random walk" through turbid tissue;
Fig. 10B illustrates a calculated light intensity spatial profile calculated
for
a light source at 4 mm depth from .the diffusion model (solid line) and with
Monte
Carlo simulations ("+" symbols);
Figs. 11A and 11 B are plots calculated from the diffusion model showing
peak intensity (11A) and spot width (FWHM) (10B) as a function of source
depth;
Fig. 12A is a surface light-intensity image from a light-emission source in
a subject, and shows the horizontal and vertical profile lines along which
light-
intensity profiles were measured;
Figs. 12B and 12C are light-intensity profiles measured along the
horizontal and vertical profile lines shown in Fig. 12A;
Fig. 13A shows spectral image plots of luciferase light-emitting sources
acquired in vitro (solid line), from the subcutaneous site shown in Fig. 13B
(squares); and from the lung sites seen In Fig. 13C (diamonds).
Figs. 13B and 13C are surface light-intensity images from animals with a
subcutaneous light-emitting source (Fig. 13B) and light-emitting sources in
the
lungs (Fig. 13C).
8
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
Detailed Description of the Invention
I. Definitions
Unless otherwise indicated, the terms below are defined as follows:
A "biophotonic function" is a mathematical function that describes a
measurable photonic output, such as a light emission profile, spectral
intensity
distribution, or integrated light intensity, in terms of photonic-source
variables,
such as source depth, size and shape, number of light-emitting cells,
wavelength-dependent scattering and absorption coefficients for specific types
of
tissue, and spectral characteristics of the light-emitter. The photonic
function
can be based, for example, on a photon-diffusion model, as discussed below, on
a Monte-Carlo simulation, or finite element analysis.
"Turbid tissue" is non-transparent tissue in a subject that has both light
scattering and light absorption properties.
"Target region°' refers to an internal subsurface region in a subject,
such
as a subsurface tissue or organ, a solid tumor, or a region of infection, that
is (i)
localized within the subject, (ii) separated from the surface region of the
subject
by turbid tissue, and (iii) preferably possesses at least one distinguishable
feature, such as a tissue or organ-specific antigen, gene or gene product,
either
as mRNA or expressed protein, a product resulting from activation,
deactivation,
or regulation, of a cell, organ, or tissue.
Light-emitting source refers to a source of light emission at the target
region. The source may itself emit visible light, such as when the target
tissue is
genetically modified to produce recombinant light-emitting protein when
specific
genes are expressed by some activation means. In preferred embodiments,
different types of light-generating proteins, such as bioluminescent proteins
(e.g.,
luciferases) and fluorescent proteins (e.g., GFP, dsRed, and YFP) are
employed.
Exemplary light-emitting sources include prokaryotic and eukaryotic cells
transformed with a light-generating protein and administered to the animal
subject, as well as light-producing cells which are intrinsically part of an
animal
made transgenic for a light-generating reporter gene.
Alternatively, the source may include fluorescent or other molecules that
can be excited by electromagnetic radiation to emit visible light. For
example, a
9
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
composition effective to cause a target to emit light may be administered to
the
subject and allowed to localize at the target location wherein the composition
will
then emit light, or cause some other moiety in or adjacent the target to emit
light.
For example, the administered composition may be a compound which activates
cells in or adjacent the target, and in-turn, causes such cells to emit light,
hence,
the localized cells are the target within the subject's body region. In yet
other
embodiments, infectious cells are administered to a subject, and the progress
and locality of the disease are determined. The target, in this embodiment, is
the
cluster of cells at a given point having a location within the body region of
the
subject. In other embodiments of the invention, a body region may have
multiple
targets, and a subject may have multiple body regions, each region potentially
having a target.
"Fluorochromes" or "fluorophores" are molecules which are excited at or
effectively near an excitation maxima, and emit or fluoresce light at longer
wavelengths. The difference between the excitation maxima of a fluorochrome,
and the emission maxima of the fluorochrome, is known as a Stokes shift. The
greater the Stokes shift for a given fluorochrome, the greater the difference
in
between the excitation spectrum, or range of wavelengths, which excites the
fluorochrome, and the emission spectrum or range of wavelengths of light
emitted by the fluorochrome when is fluoresces. Certain embodiments of the
invention employ fluorochromes with high Stokes shifts such with Texas Red,
rhodamine, Cy3, CyS, Cy7, and other deep, far red, or near infra-red (NIRF)
fluorochromes. Particularly preferred fluorochromes emit light at wavelengths
longer than about 600 nanometers. Other preferred fluorochromes will emit
light
in ranges shorter than 600 nanometers. In particular, variants of "green
fluorescent protein" (GFP), each having various emission spectra, are
particularly useful because, like luciferase, such proteins may be synthesized
by
the target or cells adjacent or associated with the target, or which localized
at the
target, for example, as in an infection. GFPs may further be administered as a
conjugate to another protein or biological material that localizes at the
target or
targets.
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
In preferred embodiments of the invention, each light-emitting source,
under known light-emitting conditions, has a known emission spectrum. Each
emission spectrum is preferably constant throughout the detection step. In
particular, the relative intensity of light emission at each wavelength of
light
remains relative to the intensity of light emission at all other wavelengths
of light
within the emission spectrum of the light-emitting source, as the overall
intensity
of the spectrum changes. Some embodiments of the invention minimally require
at least two distinguishable ranges of wavelengths of emitted light. More
preferably, particular embodiments of the invention require that each range of
wavelengths of light behave differently, as a function of the distance or
depth for
which such light travels through the body region, from other ranges as they
travel
through a selected body region of a subject. For example, the intensity of a
first
range of wavelengths of light is cut in half for every centimeter of a body
region
the light travels, whereas a second range of wavelengths of light is cut by
three
fourth for every centimeter of body region the light travels through. The
intensity
of light having both ranges of wavelengths, where each range has the same
intensity, will diminish by one half at the first range, and three quarters at
the
second range, after passing through one centimeter of tissue. By comparing the
resulting 2:1 ratio of intensities from the initial 1:1 ratio of intensities
at tissue
depth zero, a determination that the light traveled one centimeter through the
tissue can be made.
The invention may be practiced using different light-emitting sources such
as different luciferase enzymes. For example, the light emission from "blue"
bacterial luciferase (Photorhabdus luminescens) expressed on agar was
analyzed by a spectrometer that revealed a spectrum centered about 485 nm. A
"green" firefly luciferase (Photinus pyralis) expressed in PC-3M cells in PBS
solution emitted light demonstrating a spectrum centered about 570 nm. A "red"
luciferase expressed in PC-3M cells suspended in PBS solution demonstrated a
spectrum centered about 620 nm.
11
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
II Ap~~~aratus and Method
Fig. 1A is a cross-sectional view of an apparatus 20 which may be used in
investigating the location and size of a light-emitting source in a subject,
in
accordance with the invention. The apparatus generally includes a light-tight
enclosure or chamber 22 within which light-emission events from the subject
can
be detected. Light tight box 22 may be constructed as described in co-owned
PCT publication number WO 200163247, which is herein incorporated by
reference in its entirety. Briefly, the chamber is defined by back and side
walls,
such as back wall 21, and has a front opening (not shown, which can opened to
provided access to the interior of the chamber, and closed to provide a light-
tight
seal. Although light tight box 22 was designed for small animals, the
technology
can also be applied to larger mammals, including humans.
Contained within the chamber is a stage 24 on which the subject is placed
and preferably immobilized during light-emission detection. Although many of
the examples. described herein have to do with a small mammalian subject,
typically a mouse or a rat, it will be appreciated that the same principles
and
methods will apply to other animals as well as human subjects, employing
suitable scale-up in chamber size and where necessary, suitable increases in
the size and number of light emitters in the target region. The stage in the
apparatus shown is designed to be raised and lowered to a selected vertical
position within the chamber by a conventional wormscrew mechanism shown
generally at 26 and under user control. The stage position within the chamber
may be monitored by any of a number of known tracking devices that do not
require visible light signals.
The viewing optics includes to two general components. The first is a
lens assembly 23 that serves to focus light-emission events from the surface
of
the subject onto the detection screen of a photodetector 32, which is
typically a
photodetector pixel array of a charged-coupled device that can be operated in
a
cooled condition to minimize intrinsic noise level. The construction of the
lens
assembly and its optical coupling to the detector is conventional. One
preferred
CCD is a model 620 CCD, commercially available from Spectral Instruments.
12
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
The second, and optional, component of the optical system is a
wavelength filter wheel 30 containing a number of optical bandpass filters
designed to block light transmission in all but a selected range of visible-
light
wavelengths. One standard installation includes 3 filters; a shortpass filter
30a,
for wavelengths less than 510 nm (blue light filter), a midpass filter 30b
having a
bandpass in the 500-570 nm range (green light filter), and a longpass filter
for
wavelengths greater than 590 nm (red light filter). In some applications,
filters
can provide more precise bandpass, e.g., every 20 nm. The just-described
viewing optics and detector are also referred to herein, collectively, as an
optical
system.
The apparatus also includes a control unit 34. Various user-controlled
settings, such as stage height, orientation, and translation position,
bandpass
filter, and detector mode are made through a control input pane; 36 in the
apparatus.
Fig. 1 B shows a more schematic representation of apparatus 20, here
shown in a scale for use in imaging a small mammalian subject, such as a
mouse 38. The lens assembly is represented here as a single lens 28, the
filter
wheel at 30, and a CCD detector at 32. The detector output is supplied to a
computational unit 40 in the control unit computational unit 34 which includes
a
processor 40 and a storage unit 42 for data files. The content of the data
files,
and operation of the computational unit, in response to light-detection
signals
from the designs from detector 32 will be described below. The control unit is
operatively connected to a display/output device 44! such as a computer
monitor, for displaying information and images to the user.
The operation of the computational unit can be appreciated from the
method of the invention carried out by the apparatus, which will be summarized
with respect to Fig. 2, and detailed further below. The first step in the
method is
to obtain a first measured light-intensity profile. This is done by first
localizing a
light-emitting source within a subject. This may be done, as noted above, in a
variety of ways, including introducing into the subject, a gene efFective to
produce a light-emitting protein, such as luciferase, in the target region, or
localizing a fluorescent compound at the target source, both according to
13
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
methods which are known, e.g., as described in co-owned PCT applications
W099US/30078, WOOOUS/7296, W099US30080; and W097US6578, all of
which are incorporated herein by reference.
The subject so prepared is placed in the light-tight chamber, and light-
s emission form the subject surface originating from light-emission events at
the
target region are measured, digitized, and placed in the computational unit,
as at
52 in Fig. 2. The computational unit uses this data to generate a spatial
profile of
light emission, as at 54. This profile could be taken along a selected row and
a
selected column of detector elements in the detector representing profiles
along
x (horizontal) and y (vertical) axes in a plane parallel to the stage in Fig.
1A.
That is, the profile is a plot of measured intensity values along a selected
row
and column of detector elements, and represents the distribution of light
intensity
values that are focused from the subject surface onto the detector array.
Exemplary light-intensity values measured from a subject are shown in solid
lines in Fig. 3B. Fig 3A shows the emission contours, and Fig. 3B shovrvs the
vertical profile, as determined at detector pixels corresponding to those
positions.
Turning back now to Fig. 2, the spatial profile or profiles so obtained are
matched or fitted with profiles stored in a database 58 of parameter-based
photonic functions which are either empirical functions measured from light-
emitting sources of known size and depth, or are calculated from photon-
diffusion models using depth and size parameters, and optionally, other
parameters, as will be described below.
From the optimal curve match, the program can make an initial estimated
determination of depth and/or size and/or brightness, preferably all three, of
the
light-emitting source. According to a feature of the invention, the method is
now
carried out in a way that the depth and size determinations are refined by
additional data that may be in the form of additional light-intensity data,
and/or
additional i~nodeling data. For example, and as indicated in Fig. 2, the data
may
be in the form of:
(a) 2nd-view data, as at 62, meaning light-intensity data, typically spatial
profile data, obtained by viewing the subject from a second view angle.
14
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
Typically, the subject in the apparatus is tilted with respect to the optical
system
so that light-emission is obtained from a second region of the subject
surface;
(b) Spectral data, as at 64, meaning light-intensity data obtained at one or
more bandpass ranges, e.g., in a blue-light, green-light, or red-light
spectral
range;
(c) Transmittance data, as at 66, meaning predefiermined-light intensity
values obtained typically as a function of wavelength at selected locations in
a
subject that approximates the subject of interest;
(d) Tissue properties data, as at 68, meaning predetermined values of
tissue parameters, such as reduced scattering coefficient,u'S, absorption
coefficient,ua, or effective coefficient,ueff corresponding to the subject
tissue
through which light is diffusing. This data is used, for example, to refine
the
spatial profile curves generated by a diffusion model;
(e) Simulation data, as at 70, meaning predetermined light-intensity
values, typically spatial profile data, obtained by placing a light-source of
known
' intensity and size at selected locations in a model representing the subject
target
region and adjacent surface, or in an actual subject that approximates the
subject; and
(f) Total intensity data, as at 72, meaning total light intensity summed or
integrated over an entire spatial profile or an entire detector array.
The program run by the computation device receives the additional information
at 60, and uses the information to refine the depth and size determination of
the
light-emitting source, at 74, as will be considered below. Finally, the
refined
determination of source depth, size, and optionally, shape and number of cells
in
the light-emitting source, is displayed to the user, either in the form of
data
and/or one or more subject images showing the location and intensity of the
light-emitting source.
(g) Absolute calibrated light intensity, as at 73. Calibration to absolute
intensity allows one to convert counts/sec/pixel in the CCD to radiance
photons/s/cm2/sr (sr=steradians). Fig. 8A illustrates an optical assembly 100
used in the calibration method. The assembly includes, in addition to a CCD
detector 102, lens 104 (representing a lens assembly, as above), a bandpass
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
filter 106 and a low-light-level integrating sphere, such as a OL Series 425
sphere available from Optronic Laboratories, which acts as a source of a known
light intensity. The known radiance from the sphere, measured in
photons/sec/cm2/sr, allows one to calculate the calibration factor that
relates the
two numbers.
Fig. SB illustrates how absolute measured intensity is used to calculate
total number of cells in a light-emitting source. It is assumed that each cell
will
produce a photon flux ~~ photons/sec/cell. The radiance from a light-emitting
source 110 in a subject 112, as measured by detector 102, measures the
number of photons/s/cm2/sr. The measured radiance is then integrated over a
region of interest to convert to a source flux ~S photons/sec. From the known
flux values, the number of cells in the light-emitting source is readily
calculated
as ~Slep~.
III. Light-emission data and parameters
This section considers various types of light-emission data that may be
collected in practicing the invention, and discusses how the data may be used
in
photon-diffusion modeling, and/or for determining depth or size of a light-
emitting
source based on curve matching with,a model diffusion curve.
Light-intensitYspatial profile. The basic light-emission data that is
determined is a light-intensity spatial profile which has been described
above,
with reference to Fig. 3B. A measured light-intensity spatial profile shows
the
spatial distribution of light intensity emitted from the surFace region of
subject.
The light-intensity spatial profile may be obtained along a single line, along
2 or
more lines (e.g., along the x and y directions with respect to the stage
supporting
the subject). The entire 2D surface distribution of light may also be used. A
modeled spatial profile represents the predicted light-intensity spatial
distribution
calculated from a model of photon diffusion through a turbid tissue. By
matching
a measured profile with a modeled profile, an approximate depth and/or size of
light-emitting source can be determined.
Total light intensity. Total light intensity is the total light intensity
summed
or integrated over all of the detector elements in the detector. Total light
16
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
intensity can be used as a further constraint on the fitting model used to
determine source, location, and brightness.
Absolute calibrated light intensity. As described above with reference to
Fig. 8B, the absolute light-intensity flux can be used to measure total flux
from
the light-emitting source. Assuming that this source is made up of cells which
each emit an average cell flux ~~, the total number of light-emitting cells
making
up the source can be determined. This determination, in turn, can bemused to
refine the total source mass or volume determination, based on a known or
estimated cell mass ratio for the tissue source.
Spectral data. Spectral data is light-intensity data collected at particular
wavelength ranges, e.g., the blue-light, green-light, and red-light wavelength
ranges noted above. Because light is absorbed preferentially at shorter
wavelengths by subject tissue, and particularly by hemoglobin in subject
tissue,
the relative intensities of light at different wavelengths, can provide
information
about source depth and can be used in producing refined wavelength-dependent
spatial profiles, based on wavelength-dependent tissue scattering and
absorption coefficients.
Figs. 4A and 4B depict the differential effect distance or depth has upon
light that travels through the tissue of the body region, at different
wavelengths or
ranges of wavelengths. The graph shows increasing light intensity on the y-
axis
and increasing light wavelength on the x-axis. The solid line in Fig. 4A
represents the in vitro emission spectra for the light-emitting source that
correlates to a zero-depth emission spectrum which comprises a broad single
peak centered about 495 nm. The broad dashed line represents the emission
spectrum of the light-emitting source after it has passed through about one
millimeter of tissue (in vivo). At about 600 nm, the one millimeter depth
emission
spectrum develops a small peak, as the once large peak at about 495 nm begins
to diminish in relative intensify. The intermediate dashed line represents the
emission spectrum of the light-emitting source after it has passed through
about
two millimeters of tissue (in vivo). The 495 nm peak becomes lesser in
intensity,
while the 600 nm peak increases and becomes more pronounced. Lastly, the
dotted tine represents the emission spectrum of the light-emitting source
after it
17
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
has passed through about four millimeters of tissue (in vivo) where the 495 nm
peak has been further reduced.
Fig. 4B represents the emission spectrum at four millimeters depth of
tissue scaled against the zero tissue depth emission spectra. The 600 nm peak
is prominent over the once prominent 495 nm peak. In this example, it is the
ratio of the 600 nm peak to the 495 nm peak that provides information about
the
depth of the fight-emitting source.
Preferred methods of the invention employ self-scaling computations so
that the depth of the light-emitting source may be determined independent of
its
overall intensity. To achieve this end, functions are formulated that utilize
ratios
of two or more ranges of wavelengths, where each range of wavelengths travels
through tissue in a body region at different rates as a function of depth of
the
target from the surface of the body region.
Through-body transmittance. Through-body transmittance is measured
. by placing a know-intensity light-source at a selected location against or
inside a
subject, and measuring the light-intensity that is transmitted through the
subject
on multiple surfaces. The transmittance will have a strong spectral component,
owing to the strong spectral dependence of the scattering and absorption
coefficients of different tissues (see below). Typically, transmittance values
are
pre-determined using a model subject, by measuring total transmittance at
various selected locations and at each of a number of spectral ranges.
Fig. 5 shows a top view of a subject having a plurality of body regions
indicated by numbers 1-16. In this example, the subject is an immobilized
mouse, and each body region corresponds to pre-determined regions of interest.
Prior to depth determinations by the practice of the methods of the invention,
the
subject is optically analyzed to develop a data file of optical
characteristics for
each body region. In this example, each body region is placed between a
spectral light source and a spectral detector. Light transmission spectra are
obtained for each body region. The result is analogous to placing each body
region in a spectrophotometer cuvette and measuring the spectral profile of
each
region at different wavelengths and wavelength ranges, to create a data file
of
how light transmits through different body regions, and the tissues within
such
18
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
regions. The device of Fig. 1 may be used to develop similar data for
particular
ranges of wavelengths provided the device is adapted to trans-illuminate, with
respect to the detector, the subject at each body region tested. Thus, the
device
of Fig. 1 becomes, in essence, a whole animal spectrophotometer.
Fig. 6A and 6B show the empirical spectral data gathered from the
spectral analysis described immediately above. As is described below in more
detail, this data, either as whole spectra, or particular ranges within each
spectrum, can be used to create the data files for comparison with detected
intensity information, or for developing functions to compute target depth.
In a more detailed approach, the apparatus provides a whole body
scanner for scanning an entire subject, or portions thereof, which further
comprises a scanning two-color laser for scanning the body of a subject. The
scanner moves through an arc laser light of at least two colors, either
simultaneously, or sequentially. As the laser light beams upward, it
penetrates
the body of a subject placed within a rotational tube. The detector placed
opposite the laser light source is adapted to receive and measure the laser
light
as it moves through its arc path.
Tissue optical properties. There are two important optical properties of
subject tissue used in photon diffusion modeling. The first is the absorption
coefficient,ua of the tissue, which is related to the fraction of incident
light that is
absorbed per unit pathlength. As seen in Figs. 7A and 7B, which plot
absorption
coefficients for a variety of tissue types over the visible spectrum, the
absorption
coefficient is highly dependent on the nature of the tissue, and the
wavelength of
light, with each tissue showing a peak absorption in the 500-600 nm range. The
relatively low absorption coefficient above about 600 nm is consistent with
the
spectral data seen in Fig. 13A, and with the total transmittance data shown in
Figs. 6A and 6B.
The second important optical property is the reduced scattering coefficient
,~'S of the tissue, which is the fractional decrease in intensity per unit
length of
penetration of light due to large angle light scattering. As seen in Figs. 7C
and
7D, which plot reduced scattering coefficient for the same tissues over the
visible
spectrum, scattering effects can be relatively different in different tissues.
In
19
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
general, the reduced scattering coefficient decreases somewhat at longer
wavelengths.
A double integrating sphere (DIS) system is a widely used tool to measure
the optical properties of tissue or any turbid medium since it conveniently
measures the diffuse reflectance, Rd, and the diffuse transmittance, Td,
simultaneously. Fig 6C illustrates the double integrating sphere apparatus 74
used to measure these values in order to obtain optical properties such as the
absorption coefficient, Via, and the reduced scattering coefficient, ~,'S
(Prahl, S.A.,
et al., Applied Optics 32:559-568 (1993); Pickering, J.W., et al., Applied
Optics
32:399-410 (1993)). The sample 76 is placed between the two integrating
spheres 78 and 80, with internal baffles 82 positioned to prevent measurement
of directly reflected or transmitted light. The spheres are coated with a
highly
reflective material to minimize absorption of light. Light is detected with
detectors 84 and preferably analyzed using a computer 86. Sample illumination
between 400-1000nm is achieved using an arc lamp connected to a
monochromator 88, via a fiber optic cable 90 and lens 92.
From diffuse reflectance, Rd, and diffuse transmittance, Td, values, the
absorption and the reduced scattering coefficients are obtained using the
inverse
adding-doubling program (http://omlc.ogi.edu/software/iad/index.html). This
program is a numerical solution to the one speed radiative transport equation,
which describes light propagation at steady state in a scattering medium
(Prahl,
S.A., et al., Applied Optics 32:559-568 (1993)). The program is an iterative
process, which estimates the reflectance and transmittance from a set of
optical
parameters until the calculated reflectance and transmittance match the
measured values.
A characteristic of the DIS system is the fact that tissues need to be
extracted before the measurement. Therefore, it is desirable to maintain
tissue
viability under measurement conditions. Vascular drainage of hemoglobin and
tissue hydration need to be considered, especially at wavelengths where
hemoglobin and water absorption are high. Partly in view of the foregoing, a
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
preferred method of measuring tissue optical properties is a non-invasive in
vivo
measurement, as follows.
A radial fiber probe similar to that described by Bevilacqua (Bevialcqua,
-~ F., et al., Applied Optics 38:4939-4950 (1999)) may be used to measure
optical
properties non-invasively or with minimal invasiveness in vivo. Figure 6D
illustrates an experimental set up of a radial fiber probe 94. Fiber probe 94
includes an illumination source 95, at least one illumination fiber 96 and
several
(e.g., six) detection fibers 97 generally within a 1-2 mm distance. The probe
is
then place perpendicularly on a tissue of interest 99, preferably with minimal
contact pressure. The output of the detection fibers 97 is fed into a set of
corresponding detectors 98. The optical properties are spatially resolved from
intensity versus radial distance from the source. At small source-to-detector
separations, simple analytical models such as the diffusion approximations are
not well suited to describe the system therefore Monte Carlo methods are
typically used to analyze the data.
IV. Photon transport models
This section considers photon-transport models for simulating photon
emission'at the surface of a body from a light source below the surface and
moving through a turbid medium. In particular, it is desired to generate
simulated light-intensity spatial profiles based on (i) depth of the light-
emitting
source, (ii) size of the light-emitting source, and (iii) light-absorption and
light-
scattering coefficients. Since the light-absorption and light-scattering
coefficients
are dependent both on wavelength and on the nature of the tissue through which
the photons diffuse, the model may be refined to specifically consider the
nature
of the tissue and the spatial profiles at selected wavelengths.
The photon-diffusion model employed herein makes a number of
simplifying assumptions for purposes of generating an initial light-intensity
spatial
profile from which initial source-depth and source-size information can be
generated, as will be considered in Section V. The model may then be
expanded by considering additional tissue-dependent and/or wavelength
dependent information to refine the spatial profile, for purposes of improving
the
21
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
curve fit with a measured spatial profile. Alternatively, the initial depth
and size
information can be refined by other types of curve or data matching. Both
approaches for refining the initial depth/size approximation will be
considered in
Section V below.
Fig. 9A shows a photon path as it diffuses through a turbid tissue
formed of cells, such as cell 111 containing organelles 113, As seen, the cell
size, typically in the 20-30 micron size range, is large relative to the
wavelength e1
of light, which is between about 0.4 and 0.6 microns. Scattering is due to
abrupt
discontinuities («a ) at membranes and is characterized by an inverse length
scattering coefficient,us of about 10-20 mm-'. The scattering anisotropy g is
about 0.9, which gives a reduced scattering coefficient,u'S =,us (1-g), or
about 2
mrrf'. The absorption coefficient,ua (a) is between 0.01 and 1 mm' , and as
seen above, has a strong wavelength dependence. The absorption and
scattering coefficients are wavelength-dependent, but for a> 600 nm, it is
generally true that,~a «,u'S in tissue.
A quantitative description of transport of light from a target site within a
body to a light detector located adjacent a body surface may be achieved by a
solution of a radiative transfer diffusion equation of light with appropriate
boundary conditions (e.g., R.C. Haskel; et al in "Boundary conditions for the
diffusion equation in radiative transfer", J. Optical Soc Am., 11A:2727
(1994)).
One approach, for example, uses an extrapolated boundary condition shown
schematically in Fig. 9B in which optical fluence vanishes at a planar surface
that
is displaced from the physical boundary at a distance zb that takes account of
the
Fresnel reflection of light at the physical boundary.
As a first approximation, the body in which the light-emitting source is
located is represented as a semi-infinite, homogenous turbid medium that is
both
scattering and absorbing. The approximation is illustrated in Fig. 9B which
shows a light-emitting point source 114 in slab 116 a distance d below the
slab
surface. The diffusion equation is
DOz~~y') = f~n~~~') - S(~') (1 )
22
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
and Fick's law is
j(r) =-D0~(r) (2)
where ~ is the isotropic fluence (watts/m2), j is the small direction flux
(watts/m2),
S is the power density (watts/m3), r is the radius, and
_ 1
D 3L(1-g)A~s +,~a~ (3)
The Green's function solution for a point source P (watts) is
~(j') = 4 Dr exp( ~e~r) (4)
where ,ue~. = 3,Cla ,Lls + ,Ch and ,us = (1- g),Cls .
The solution to the diffusion equation in slab geometry using the
extrapolated boundary condition is obtained by summing contributions from the
source 114 plus image source 118 (as shown in Fig. 9B), resulting in the
following equation for radiance at the surface (viewing perpendicular to the
surface) for a point source P (watts) at a depth d.
exp(-,ue~.r~) exp(-f.~e~.r2)
ri rz
d 1 d + 2zb
4~c 4TCD j.z peg' + ~, exp(-,ueffr;) + r z (5)
+3D 1 ' z
1
x ,ue~- +- exp(-,ue~ra)
rz
where r = p2 +d2 , r2 = p2 +(d +2Zb)2 p is the radius on the slab
surface and
_1+~~f ~
zb 1-Reff 3(,ua +f~s)
and Reff is the effective reflection coefficient which is about 0.43 for
tissue.
23
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
Plotting surface radiance as a function of p gives the plot shown in the
solid line in Fig. 10B, which has the same derivation as the light-intensity
profile
plot shown in dotted lines in Fig. 3B.
The spatial profile curve for a light-emitting point in a turbid medium was
also calculated using a much more computationally intensive Monte Carlo
simulation, which follows each photon through a random walk, as illustrated in
Fig. 1 OA. The Monte Carlo simulation, shown in "+" symbols in Fig. 10B
closely
matches the spatial profile calculated from the diffusion equation.
To illustrate how the diffusion equation may be used to model photon
diffusion through various tissues and a various depths, the spatial profile
was
calculated for various values of,~eff , corresponding to ga and,u'S values of
2.0 and
cm ~, respectively, for,~eff = 11 cm-~; 0.4 and 10 cm ~, respectively,
for,~eff =
3.5 cm~~ , and 0.05 and 5 cm ~, respectively, for,~eff = 0.87 crri ~. The
three
values of for,~a correspond roughly to tissue absorption for blue, green, and
red
15 wavelengths.
The plots in Figs. 11A and 11 B show that intensity decreases and spot
width increases with increasing depth, and that large values of,~eff result in
large
attenuation of peak intensity with narrower spot width.
20 V. Determining Depth and Size information
In practicing the method of the invention, the subject is initially treated to
localize light-emitting molecules at a target source, as detailed above. The
subject is then placed at a selected location within the light-tight chamber
of
apparatus 20 (appropriately scaled for subject size) and the optical system is
adjusted to measure light emission events from the source at the surface
region
of the subject between the light-emitting source and the detection optics.
With
the subject immobilized at a selected optical perspective, a light-intensity
spatial
profile of surface light emission is obtained, also as described above. Fig.
3A is
a surface map of measured light intensity from a subsurface light source,
which
was subsequently used to generate the spatial profile shown in Fig. 3B.
The spatial profiles) are then fitted with parameter-based biophotonic
functions that relate light-emission intensity to depth of a light-emitting
source
24
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
(and in some cases, source size) to obtain an initial determination of source
depth (and in some case, source size). One preferred biophotonic function is
that derived from the simplified diffusion model discussed in Section IV
above, in
which light intensity from a point source or source having a defined spherical
or
ellipsoidal volume is calculated as a function of source depth, fixed
scattering
and absorption coefficients, and surface distance r from the source.
Conventional non-linear least-squares curve fitting techniques, such as
Levenberg-Marquardt ("Numerical Recipes in C", Press et al., eds, Cambridge
Press, NY, 1989). are suitable for the curve fitting. Curve fitting can be
done
using a single 1-dimensional ("1-D") profile, as shown in Fig. 3, a plurality
of such
profiles, or the entire 2-D spatial distribution (e.g., as shown in Fig. 12A).
Curve-
fitting calculations to the data shown in Fig. 3B (dashed line) indicate a
depth of
2.7 mm, which compares well with the estimated actual depth of 2.2 mm.
Figs. 12A-12C give another example of how depth and size information
can be determined from spatial profiles. The light-emitting source here is a
subcutaneous elliptical tumor having actual dimensions in the horizontal and
vertical directions of 2.7 mm and 3.2 mm, respectively, and a tumor thickness
of
1.5 mm. The horizontal and vertical profiles were (Figs. 12~ and 12C,
respectively) were fitted with curves generated for an ellipsoid light-
emitting
source having horizontal and vertical dimensions of 1.3 and 2.4 respectively,
a
thickness of 1.5 mm, and a depth of 0.4 mm.
The source depth (and optionally, source size) determination from the
initial curve fitting can be refined, and source size approximated, by
employing
additional data whose effect is to refine the parameters of the biophotonic
function. The nature of the additional data, and the manner in which it
refines
the parameters, e.g., depth and size, of the biophotonic function, providing
refined depth and source-size information, will now be considered for each of
the
information types outlined in Section II above, with respect to Fig. 2.
2nd-view information. To obtain 2nd-view data, the subject is rotated
with respect to the viewing and detection optics for viewing the light-
emitting
source from another perspective. From this second perspective, a second light
intensity profile is obtained, providing a second depth determination with
respect
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
to a second subject surface region. By determining the intersection of the
source depths at two different surface locations, a more accurate depth and/or
source size can be determined. The more views that are obtained, the more
accurately depth and source size can be determined.
B. Spectral data. A target's depth within a body region may be
determined, or the target depth refined, by (i) causing the target to emit
light at
two or more wavelength ranges, and (ii) comparing the differences in each
range's light transmission through the body region, at a surface of the body
region, provided that the light in each wavelength range transmits through the
body region differently from the light in other wavelength ranges as a
function of
depth.
Typical spectral curves (representing total light intensity at the indicated
wavelengths) as a function of source depth are shown in Fig. 4A, discussed
above. As seen particularly in Fig. 4B (where the peaks are scaled to the in
vitro
spectral curve), the ratio of peak height at above 600 nm, e.g., 620 nm, to
that
the in vitro (zero-depth) peak height, e.g., around 500 nm, increases
dramatically
as a function of depth. Thus, by generating a spectral curve and determining
the
ratio of peak heights, typically above and below 600 nm, e.g., 620nm: 500 nm,
accurate depth information based on the wavelength-dependent loss in intensity
in turbid tissue, can be determined.
A useful formula for determining approximate depth from spectral
measurements can be derived from Eq. 4 as follows:
d ~ ~OD~ ~ ~2D2~
~e~j'2 Jeff 1
where d represents depth, ~ represents the measured light intensity,,ue~
represents the empirically determined effective coefficient of attenuation,
and D
is the empirically determined diffusion coefficient. The subscripts 1 and 2 in
the
above formula refer to two separate wavelengths at which measurements are
made.
26
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
Using the above formula, depth may be determined from two or more light
intensity measurements at different wavelengths or ranges of wavelengths. In
one embodiment, the depth may be obtained by executing the following series of
steps: (1 ) Image the bioluminescent cells in vitro as well as in vivo (in the
animal) at two wavelengths; (2) Quantify the images by measuring, e.g., the
peak intensity or the average (integrated) intensity for each image; (3)
Calculate
the ratio of the in vivo image data to the in vitro image data at each
wavelength;
and (4) Calculate the depth using Equation 7 and an effective scattering
coefficient p,effobtained, e.g., from tissue property measurements. An
application
of this approach is illustrated in Example 1.
In one embodiment, it is preferable to have a significantly different value
for each ~e~. Animal tissue provides such a difference in Jeff, due largely to
the
presence of hemoglobin, which has a large absorption peak just below 600 nm,
and relatively low absorption above 600 nm.
Alternatively, the additional spectral data may include one or more spatial
profiles taken at one or more selected .wavelengths or wavelength ranges. The
profiles) are then compared with model intensity functions generated, for
example, from a photon diffusion model above, using wavelength-specific values
for the absorption, scattering, and/or effective coefficients, as described
above
with respect to fig. 3B.
In still another application, intensity values can be measured at discrete
wavelengths, spaced every 20 nm. The spectral plots in Fig. 13A show green
luciferase in cells in PBS solution, and spectral curves for green luciferase
localized subcutaneously at a site indicated in Fig. 13B and in the lungs as
shown in Fig. 13C. All the curves are normalized to a value of 1 at 700 nm.
The
ratio of intensities at 560nm to 620nm is indicative of depth.
As noted above, a related aspect of the invention provides a method of
determining the depth of a light-emitting source in a subject. In practicing
the
method, the light emission intensity from the subject at two or more different
wavelengths between about 400 and about 1000 nm. The depth of the light-
emitting source is determined using the measured light-emission intensities,
and
27
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
information related to the optical properties of the subject, e.g.,
coefficients of
attenuation and diffusion in the subject.
Information relating to the optical properties, .e.g., coefficients of
attenuation and diffusion in the subject may be obtained by direct measurement
of the coefficients in material that is the same or similar to that of the
subject.
This information, combined with measured intensities at two wavelengths in
vitro
(zero depth) and in vivo (the depth to be determined) can be applied to
Equation
7 to determine the depth of the light source, as described above, and as
illustrated in Example 1.
Alternatively, optical-properties information may be obtained indirectly, by
determining, at two or more different wavelengths, light intensities from a
light-
emitting source located at each of a two or more of a plurality of depths in
tissue
or material corresponding to that of the subject. The desired depth
determination is then made by matching the measured light intensities, e.g.,
the
ratio of the intensities at the different wavelengths, with the light
intensities
determined at each of the plurality of depths. In a more specific approach,
the
spectral profile of light intensities from the light-emitting source may be
compared (matched) with each of a plurality of spectral profiles of light
intensities
from a light-emitting source at each of a plurality of depths.
C. Through-body transmittance data. Through-body transmittance data
is obtained as described above, and illustrated in Figs. 5, 6A and 6B for a
small-
animal subject. As noted above, the information provides a wavelength
dependent light-transmission value through a selected tissue or tissues of a
known thickness. Predetermined transmittance data from a number of selected
locations in a subject can be used to estimate average whole-body scattering
and absorption coefficients at the different subject locations, for purposes
of
refining the modeled spatial profile, either over the entire visible spectrum,
or at
selected wavelengths, used for curve fitting to the measured spatial profile.
D. Tissue properties data. The tissues property data, such as that shown
in Figs. 7A-7D, typically includes wavelength dependent absorption and
scattering coefficients for each of the major body tissues. This data is used,
for
example, in the above model of photon diffusion, to refine the spatial
profiles that
28
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
would be produced by light transmission through a given tissue, at a given
wavelength, typically in the red wavelength. Thus, for example, if the light-
emitting source in the subject is muscle, and the spatial profile is taken in
the red
wavelength, a refined spatial profile generated with the tissue-specific and
5' wavelength-specific absorption coefficient will provide a refined spatial
profile
curve for curve fitting with the measured curve.
E. Simulation data. In another embodiment, a test point or otherwise
known-shape light-emitting source may be introduced into a block or slab that
simulates turbid tissue. The block can be prepared with various scattering and
diffusion coefficients, and various shapes to simulate light-emitting source
conditions in a subject. Upon placement of the test point, spectral profiles
may
be taken from the outer surface of the body region where the test point was
placed. This data then is correlated with the actual depth and location of the
test
point. By moving the test point from point to point within the block, a series
of
spectral measurements can be made. From this series of spectral
measurements, a data file can be assembled to. model the spectral responses of
different regions within the subject.
F. Integrated light intensity
Summed or integrated light-intensity, as noted above, refers to the light-
intensity summed over all or some defined area of the detector array. The
integrated light intensity can be compared with the integral of Equation 5 to
provide yet another estimate of source depth and brightness. This information
can be used in conjunction with the profile information. The integrated light
intensity can also be calculated for multiple wavelengths.
G. Calibrated-intensity data. In the absence of calibrated intensity data,
intensity measurements may vary from camera to camera, and measured values
will depend on variable such as field of view, binning, time, and f stop.
Absolute calibrated intensity (Section III above), allows one to refine the
depth-of-source determination, based on curve fitting to true peak value, and
to
estimate the number of bioluminescent cells in tissue, as discussed in Section
III
above.
29
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
In still other embodiments, the invention provides for the integration of
other imaging data with the intensity and spatial distribution data to produce
greater detail three-dimensional maps of the subject and target located
therein.
For example, compiled differential light transmission data, such as disclosed
above, may be interlaced with coordinate systems derived from other three
dimensional imaging systems such as stacked slices of MRI images forming a
three dimensional digitized "image" or coordinate system. Such a coordinate
system may then be used to better illustrate the anatomy of the subject for
which
a target is being identified. In the case where the subject is a laboratory
animal
such as a rat, such animals are highly uniform in structure from one animal to
the
next for a given strain. Consequently, a coordinate data, spectral data, and
spatial intensity pattern files may be developed by a commercial vendor, and
sold for use with simplified detection units such as shown in Fig. 1. If model
three-dimensional information is supplied as a data file, then user units do
not '
need to be equipped for three-dimensional scanning and mapping. The two-
dimensional images contemplated in the methods and apparatuses disclosed
above may be combined with the three-dimensional data files provided by the
vendor to yield complete three dimensional information about the size, shape,
depth, topology, and location of a target within the subject.
The following examples illustrate, but in no way are intended to limit the
present invention.
EXAMPLE 1
CALCULATING DEPTH OF A LIGHT-EMITTING OBJECT USING TWO-WAVELENGTH
SPECTRALIMAGES
The spectral information shown in Figure 13A was used to calculate the
depth of luciferase-labeled cells in an animal that received a sub-cutaneous
injection of the cells (Fig. 13B) and an animal having labeled cells in its
lungs
(Fig. 13C). The analysis was performed using data at wavelengths equal to 600
nm and 640 nm using the following steps: (i) the bioluminescent cells were
CA 02447262 2003-11-13
WO 02/093143 PCT/US02/16006
imaged in vitro as well as in vivo at 600 nm and 640 nm; (ii) the images were
quantified by measuring the average intensity for each image; (iii) the ratio
of the
in vivo image data to the in vitro image data were determined at each
wavelength; and (iv) the depth was calculated using Equation 7 and an average
effective scattering coefficient ~.e~ estimated from the absorption and
scattering
coefficients in Fig 7.
For the subcutaneous cells, the ratios of in vivo to in vitro intensities
(relative intensities, or ~) are 0.35 and 0.75 at 600 nm and 640 nm,
respectively.
For the lung signals, these same ratios are 0.05 and 0.47. Using ~a=0.25 mm-'
and ~,S'=1.0 mm'' at 600 nm and ~,a=0.05 mm-~ and ~S'=1.0 mm-' at 640 nm
results in Jeff=0.97 mm-~ at 600 nm and 0.4 mm-' at 640nm. The values of the
diffuse coefficient D are 0.27 mm and 0.32 mm at 600 nm and 640 nm
respectively. Substituting these numbers into Equation 6, reproduced below,
~e,/J'2 ~effl
(in this case, subscript 1 refers to 600 nm and subscript 2 refers to 640 nm),
results in d = 1.6 mm and d = 4.0 mm for the subcutaneous and lung depths,
respectively.
Although the invention has been described with respect to particular
embodiments and applications, it will be appreciated that various changes and
modification can be made without departing from the invention.
31
